The study was underpinned by a conceptual model of combining the technology acceptance theory, diffusion innovation theory and a technology infrastructure viability analysis. According to Ward (2013), the uniqueness of this combination is that three of the models have similar components, but each has a different emphasis. Although their complexity has been amplified over recent years, their predictive power is still maintained. The conceptual model has been designed so that a greater understanding of the subjects in review can be optimally analysed.
2.5.1 The Technology Acceptance Model (TAM)
TAM has been proposed by Davis (1985), and it provides an indicator of the main drivers that influence user acceptance of technology based innovations (illustrated in Figure 2). This theory enables system designers and implementers to evaluate proposed new systems prior to these systems being used in a fully-fledged manner. Davis (1985), asserts that the overall attitude towards using a given system is hypothesized to be a major determinant of whether a person actually continues using the system or not.
Figure 2-The Technology Acceptance Model as proposed in Davis (1985)
28 As indicated in Figure 2 the TAM constructs consists of “Attitude towards using” as a function of two major factors underpinning TAM, Perceived Usefulness (Thangaraj et al.), and Perceived Ease of Use (PEOU). PU is the degree to which an individual believes that using a particular system would enhance his or her job performance and PEOU is the degree to which a person believes that using a particular system would be free from effort (Davis, 1985). Davis suggests that PEOU influences PU which directly influences the intention to adopt and actually use a system. According to Venkatesh et al. (2003), TAM informs an individual’s decision to either adopt or reject a technology.
In this study, the TAM model was also chosen based on the impact it had on similar studies that were once conducted. An example is a study that was conducted by Hu et al.
(1999). The objective of the research work was to examine the applicability of the TAM in elucidating physicians’ decisions to accept telemedicine technology in the health-care context.
The results suggested that TAM was able to provide a reasonable depiction of physicians’
intention to use telemedicine technology. The user group, the technology and the organizational context were all new to IT adoption research. The study addressed a practical investigation of a technology management system that resulted in millions of dollars invested by healthcare organizations in developing and implementing telemedicine programs in recent years. The TAM model’s overall fit, explanatory power, and the individual causal links that it postulates were evaluated by examining the acceptance of telemedicine technology among physicians practicing at public tertiary hospitals in Hong Kong (Hu et al., 1999).
According to Bradley (2009), the major limitation of the TAM study, is the self-reported usage component, this means that the TAM theory does not explain how the actual system usage is really influenced, but relies on the research subject to indicate system usage. Lee et al. (2003), also concurs that the Self-reported usage is the most commonly reported limitation of the TAM theory. Instead of the model measuring actual usage, the study has to depend mainly on self-reported use, thereby assuming that self-reported usage successfully reflects actual usage. However, self-reported usage is known to be subject to the most common method bias, which exaggerates and distorts the relationship between dependent and independent variables. The other most mentioned limitation of the using the TAM theory is the tendency to examine only one information system with the same group of subjects on a single task at a
29 single point of time, thus levitating the generalization problem of any single study. The use of a homogenous set of subjects also weakens generalizability of the findings. Considering the user’s intention and perception can change over time, it is important to measure these quantities at numerous points of time. The cross-sectional study’s major weakness is that it cannot deduce the causality of the research results. The majority of the studies with lower variance explanations do not consider external variables other than original TAM variables.
A few explanations of variance are referred to as a major problem of TAM studies. Other suggested limitations of TAM studies includes single measurement scales, relatively short exposure to the technology before testing, and self-selection biases of the subjects (Lee et al., 2003).
Therefore a better approach would have been to employ an independent measure of actual use. In the context of the current study, the issues of current usage of IoT does not create too much of a controversy because currently there are aspects of IoT technology that is used in public hospitals in Zimbabwe. These include heart rate and blood pressure monitors.
The usage of this type of technology is however not widespread.
2.5.2 Diffusion of Innovation (DOI) Theory
DOI theory has been pioneered by Rogers (1995), and explains how innovations are adopted in a society. He defines an innovation as an idea, behaviour, or object that is perceived as new by its audience. According to Rogers (1995), the adoption of an innovation is reliant on specific characteristics of the innovation that determine whether potential adopters of the innovation adopt the innovation immediately, at a later stage or not at all. These characteristics of the innovation are illustrated in Figure 2.
According to Rogers et al. (2005), diffusion occurs in multifarious systems where networks connecting system members are multiple, overlapping and complex. Diffusion occurs mostly in heterogeneous zones, such as transitional spaces where adequate differentiation among network members comes to obtain a particular idea or solution. Such heterogeneous connections, which comprise the innovation-diffusion system, occur among innovators and other engaged members of target populations who, in Rogers’s original formulation, are called
30
“cosmopolites.” Cosmopolites are locally networked system members with heterogeneous weak ties to outside systems (Rogers et al., 2005).
As indicated in the Figure 3, there are five attributes that make up the Diffusion of Innovation theory. These are Relative advantage, Compatibility, Complexity, Trialability and Observability.
Figure 3 - Diffusion of Innovation Model (Rogers (1995))
Relative advantage is a reference to the perceived technical superiority that an innovation introduces so that it is an improvement over a previously used technology/innovation. Compatibility is the capacity for the innovation to be aligned with existing human skills, values, and work practices of potential adopters. Complexity ascertains whether an innovation is relatively challenging to understand and use. Trialability measures if the innovation can be experimented with on a trial basis without excessive effort and cost and without incurring too much of an influence to the domain of use; it may be implemented in parts and still provide a net positive benefit or its implementation may be reversed without incurring too much damage within the context of use. Observability is a reference to the benefits and results of the innovation`s use which can be simply communicated and
31 experiential to others. Rogers argues that diffusion of innovation explains the process that occurs as people adopt a new idea, product, practice or philosophy.
Lyytinen and Damsgaard (2001), Identifies six limitations of the DOI theory. Stating that technologies are discrete packages developed by independent and neutral innovators Therefore, technologies diffuse in a homogenous fixed social ether called diffusion arena, which is separate from the innovation locale; Diffusion rate is a function of push and pull forces;
Adoption decisions are dependent on available information, preference functions and adopter's properties; Diffusion traverses through distinct stages, which exhibit little or no feedback and Time scales are relatively short and the diffusion history is not important.
The first limitation as stated by Lyytinen and Damsgaard (2001) is that the DOI it does not consider that technologies are not discrete packages. DOl research associates an innovation with unique and measurable features (Hai, 1998, Rogers et al., 2005). With this type of definition, numerous difficulties emerge. Firstly, it is not clear whether the feature list is complete and covers all features that affect adopter's behaviour. For example, why technical style or elegance does not appear in the lists though studies in the history of technology demonstrate the opposing (Hughes, 1987). Secondly, why all technological innovations should be categorised with the same set of attributes. Thirdly, what characters play these different characteristics at different stages of diffusion? For example, compatibility may mean different things for the late and early adopters. Fourthly, the assumption disregards the socially constructed nature of big technological systems (Lyytinen and Damsgaard, 2001).
The second limitation as identified by Lyytinen and Damsgaard (2001), is that the DOI does not consider that technologies do not diffuse in a homogenous and fixed social ether. In the DOl theory, interactions between technology, adopters and suppliers are expected to occur in a relatively homogeneous space. The assumption is that the technology diffuses in this ether through the effects of these three "forces". With complex technologies however, the diffusion arenas are neither fixed nor homogeneous, instead, institutional measures technological, business context and economic constraints restructure these arenas. Therefore, in analysing complex systems’ diffusion it was found essential to use institutional concepts to dynamically draw the boundaries of the diffusion space to understand what the studied processes were like.
The institutional perspective helps focus on regimes and institutional measures that are
32 involved in defaming the mandate and scope for the diffusion course. Powerful institutional changes can fundamentally affect the speed and progression of any diffusion process by restructuring its boundaries, redefining involved entities and changing incentives (Lyytinen and Damsgaard, 2001).
Lastly among other limitations is that diffusion rate is not solely a function of push and pull forces. The DOl theory integrates two additional modes of explanation: the demand-pull and the supply-push theories (King et al., 1994). Supply-push theories estimate that specific functions of the innovation cause the technology diffusion like its functionality, or the values that enable its use. The demand-pull theories explain a technology systems diffusion by a growing demand for organizational coordination. Companies need to improve their internal operations, and change their market position by relating technical knowledge (Porter, 2008).
Several IS studies have considered both forces simultaneously, unfortunately the predictive power of the theory has been low and the results confounding (Hai, 1998)
Given these limitations Lyytinen and Damsgaard (2001), believes that armed with theoretical guidelines listed below, DOl researchers will have a upper likelihood of providing faithful explanations of the diffusion of networked and complex innovations. As a way forward it is necessary to consider the following issues while studying multifaceted networked technologies. First, to pursue to understand the networked, local complex, and learning rigorous features of technology. Secondly, seeking to understand the vital role of market making and formal structures in shaping the diffusion arena. Thirdly to focus on essential process features and all key players in the diffusion arena. Forth, to develop a multi-layered theories of diffusion that factor out structures between different layers and locales. Lastly, to use other theoretical perspectives that help cover analysis beyond questions of efficient choice.
2.5.3 Technological Infrastructure
The final construct in the conceptual model is the critical role of the technological infrastructure. According to Vashi et al. (2017), IoT technology architecture consists of 5 main layers just like the Open Systems Interconnection (OSI) model. To determine the success of the adoption of an IoT in the medical healthcare sector, it is essential to ensure that the technological infrastructure that enables implementation of IoTs are available to enable the
33 deployment of the IoT strategy. Figure 4 illustrates the key elements that define the general architecture and characteristics of a typical IoT device.
IOT Infrastructure
Figure 4-The IoT Infrastructure Architecture (Ullah et al., 2016)
As shown in Figure 4, the IoT model consists of the Perception layer, Network layer / Communication Layer, Middleware layer, Application layer and the Business layer. In these layers are technologies, hardware, software and protocols that create a technological ecosystem which enables the existence of the IoT functional architecture. Ullah et al. (2016), provide details of the six technology layers as required for the deployment of an IoT strategy in the medical healthcare sector. These are:
The Perception layer that consists of the medical monitoring devices that are used to collect data that has been transmitted from the patients
Remote sensors and actuators such as RFID tags and bio patches that provide a source of data transmission
The Network or communication layer ensures data security at access points where data is collected from the perception layer and is transferred to respective central systems via the use of Wi-Fi, 2g or 3g technology.
34
The middleware layer ensures a seamless integration of protocols and different service management systems to process, retrieve and compute data.
The Application Layer is responsible for inclusive applications management based on the processed information in the Middleware layer. These are IoT smart health applications used by practitioners and patients.
The last layer is the business layer; after data has been gathered information is now fused to produce analytics, for example, the use of decision support systems is used at this stage.
According to Jara et al. (2013), from a hypothetical perspective the perception layer, middleware layer and the network layer should ensure that the IoT infrastructure is scalable, interoperable, secure, robust, enhances economies of scale and mobility. The compatibility of technologies used in the lower two layers would enable greater chances of success to ensure ubiquitous integration in the health care systems (Jara et al., 2013). On the other hand, the application and business layers should ensure consistency in robust management of content access (security related), identity management and mobility management to enable greater success chances of the supported infrastructure.
Ray (2017), identified key hindrances that may affect the implementation of an IoT healthcare system. The first challenge is to do with standardization, considering that IoT based e-health care solutions are still in the emerging stages of development, current solutions do not conform to regulations and specific rules. This includes interoperability issues that need to be taken immediately by researchers by collaborating together for example. The second challenge that may affect IoT rollout in healthcare is the quality of service. As e-health care services require reliability, rigorous and maintainability of the system hence it should be secured that no connection, delay or data loss occurs hereby improving the quality of service.
In case the case of system failure it has to be ensure that redundant services should promptly be availed to avoid disruption. The third hindrance involves issues to do with ecological aspects. Full-fledged IoT e-health care services shall need a solution such as low cost bio medical sensors that can be easily worn or implanted into human body. The challenge is without a regulatory effect, companies that manufacture may end up risking end users as the race towards creating cheaper products. Ray (2017), suggests that the regulatory bodies such as World Healthcare Organisation and governments should provide policies and guidelines of
35 manufacturing of sensors, disposal practices, and usage pattern. Therefore IoT can be a positively disruptive for providing e-health care solutions by incorporating its technologies.
Wearable medical monitoring devices and Internet of Things clouds shall become a trend of future innovations in the health sector in coming years will severely impact upon the healthcare perceptions making it safer, smarter, usable, and economic for everyone. It is of key essence that research be done to evaluate on IoT architectures beneficial for dissemination of seamless and smart e-health care services (Ray, 2017).
2.5.4 The Conceptual Framework
The study was underpinned by a conceptual model by combining the technology acceptance theory, diffusion innovation theory and a technology infrastructure viability analysis. A significant outcome of this study is the suggestion that technology adoption needs to be studied from the perspectives of the technology as well as the social context in which the technology will be used. Further insight is provided by Sanson-Fisher (2004) who used the DOI model to understand why some clinical innovations are adopted swiftly and others tend to meet a lot of resistance. Some clinical behaviours may be adopted relatively easily because of the nature of the behaviour itself, while others may involve a complex interplay between social systems, communication style and the decision-making process.
The afore-mentioned cases provide an argument for the adoption of a theoretical model that has sufficient capacity to integrate the social and technical aspects of the adoption of innovation in the medical field. Both the technical and social dimensions of technology adoption in the medical field are prevalent in TAM and DOI theory. However, to supplement for the lack of representation from the technical dimension of the study, a specific reference to the technical infrastructure is incorporated into a conceptual model that is envisaged to be an ideal fit for the study’s purpose. This conceptual model that will underpin the study’s empirical phase is presented in Figure 5.
36 Figure 5- The Conceptual Framework
Figure 5 indicates the study’s’ proposed conceptual framework of the combination of 3 constructs that are designed to measure the social and the technical capacities of the study.
Based on the arguments illustrated the individually the separate theoretical models lack full representation to explain the predictive power of the study’s objective. However synthesised they offer a stronger robust framework to determine the study’s key questions. The TAM theory attempts to determine social factors that influence user adoption. Whereas the DOI attempts to predict key factors that influence adoptability of an innovation. The Technical Infrastructure attempts to examine the feasibility of a technology operating in a particular environment given its architecture.